Energy storage management system based on cloud deployment control

Abstract

The invention discloses an energy storage management system based on cloud deployment control. A preprocessing module in the energy storage management system is used for preprocessing acquired historical electric energy data and historical weather data; the weather prediction module is used for carrying out weather change prediction based on the historical weather data; the power utilization analysis module is used for performing power utilization load change trend analysis based on the historical electric energy data; the demand analysis module is used for determining real-time demand based on the topological graph of the power utilization system; the demand stabilizing module is used for carrying out demand stabilizing analysis based on the real-time demand, the weather change prediction data and the electrical load change trend data; and the energy distribution module is used for determining an energy management strategy based on the demand stabilization analysis data and the real-time demand so as to control energy distribution of the energy storage equipment. According to the invention, energy distribution with a better effect can be realized, excessive charging or discharging of the energy storage equipment is avoided, the peak power is reduced, and the power utilization cost is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and particularly to an energy storage management system based on cloud deployment control. Background Art

[0002] With the development of technology, the power demand has been continuously climbing. More and more industrial and commercial or household users choose to install new energy storage devices to store electric energy and control the energy distribution of the energy storage devices to save electricity costs and relieve the pressure on the power grid. Among them, weather change analysis and analysis of changes in electricity load are important steps in energy distribution. At present, it is often achieved through data statistics by relevant personnel. However, this method is too dependent on the professional qualities of relevant personnel, unable to guarantee the accuracy of the analysis results, and the efficiency is not high at the same time, resulting in the inability to effectively control the electricity demand and difficult to meet the needs of modern complex power systems. At the same time, there is a lack of demand flattening analysis in the current energy management of energy storage devices. The lack of demand flattening analysis will not be able to effectively reduce the peak value of current demand, resulting in low resource utilization rate of energy storage devices and large load fluctuations, making the energy management of energy storage devices fail to achieve the expected effect and difficult to improve the electricity utilization efficiency. Summary of the Invention

[0003] The purpose of the present invention is to overcome the deficiencies of the prior art. The present invention provides an energy storage management system based on cloud deployment control, which can avoid excessive charging or discharging of energy storage devices, reduce peak power and lower electricity costs.

[0004] To solve the above technical problems, the present invention provides an energy storage management system based on cloud deployment control. The energy storage management system includes:

[0005] A preprocessing module for preprocessing the historical electric energy data and historical weather data of a target electricity-consuming object based on a cloud management system to obtain preprocessed historical electric energy data and historical weather data;

[0006] A weather prediction module for predicting weather changes based on the preprocessed historical weather data to obtain weather change prediction data;

[0007] An electricity consumption analysis module for analyzing the trend of changes in electricity load based on the preprocessed historical electric energy data to obtain electricity load change trend data;

[0008] A demand analysis module for determining the real-time demand in the current period based on the electricity system topology diagram of the target electricity-consuming object;

[0009] A demand flattening module for performing demand flattening analysis based on the real-time demand, weather change prediction data and electricity load change trend data to obtain demand flattening analysis data;

[0010] An energy distribution module, configured to determine an energy management strategy based on the demand flattening analysis data and the real-time demand in combination with the demand management objective function, and control the energy distribution of the energy storage device based on the energy management strategy.

[0011] Optionally, the preprocessing module is configured to perform data cleaning and data integration on the historical power data and the historical weather data to obtain the preprocessed historical power data and historical weather data.

[0012] Optionally, the weather prediction module includes:

[0013] A data encoding unit, configured to encode the preprocessed historical weather data in a latent space to obtain target historical weather data;

[0014] A feature extraction unit, configured to generate meteorological element features based on historical meteorological data;

[0015] A model construction unit, configured to obtain weather label data of each observation point, and construct a weather change prediction model based on the weather label data and the meteorological element features;

[0016] A change prediction unit, configured to perform weather change prediction using the target historical weather data based on the weather change prediction model to obtain weather change prediction data.

[0017] Optionally, the feature extraction unit includes:

[0018] A grid division subunit, configured to perform spatial grid division on the historical meteorological data of each observation point to obtain the spatially grid-divided historical meteorological data;

[0019] A feature acquisition subunit, configured to perform feature compression and activation on the spatially grid-divided historical meteorological data to obtain corresponding weight coefficients, and generate meteorological element features based on the weight coefficients in combination with the spatially grid-divided historical meteorological data.

[0020] Optionally, the power consumption analysis module includes:

[0021] A fuzzy clustering unit, configured to perform fuzzy clustering analysis on the preprocessed historical power data to obtain a fuzzy clustering analysis result;

[0022] A change trend unit, configured to obtain a power consumption load prediction curve, and perform power consumption load change trend analysis using the power consumption load prediction curve based on the fuzzy clustering analysis result to obtain power consumption load change trend data.

[0023] Optionally, the fuzzy clustering unit includes:

[0024] A matrix generation subunit, configured to generate a historical data vector according to the preprocessed historical power data and determine a fuzzy grouping matrix of the historical data vector;

[0025] A result acquisition subunit, configured to perform target membership degree analysis and target grouping feature analysis on the historical data vector according to the fuzzy grouping matrix, obtain the target membership degree and the target grouping feature, and perform fuzzy clustering analysis based on the target membership degree and the target grouping feature to obtain a fuzzy clustering analysis result.

[0026] Optionally, the demand analysis module includes:

[0027] An impact analysis unit, configured to obtain the power change law based on the preprocessed historical power data and perform demand prediction impact analysis based on the power change law to obtain demand prediction impact analysis data;

[0028] A demarcation point unit, configured to determine the electricity property right demarcation point based on the power consumption system topology diagram;

[0029] A real-time demand unit, configured to determine the real-time demand of the current period based on the electricity property right demarcation point, using the demand prediction impact analysis data in combination with a linear programming model.

[0030] Optionally, the real-time demand unit includes:

[0031] A first real-time demand subunit, configured to determine the first real-time demand of the current period according to the electricity property right demarcation point in combination with the slip method;

[0032] A demand calculation subunit, configured to determine the second real-time demand according to the demand prediction impact analysis data in combination with a linear programming model, and calculate the real-time demand of the current period according to the first real-time demand and the second real-time demand.

[0033] Optionally, the demand smoothing module includes:

[0034] An adjustment data unit, configured to determine power adjustment data based on the real-time demand, weather change prediction data, and power consumption load change trend data;

[0035] A priority analysis unit, configured to perform adjustable load priority analysis and interrupted load impact priority analysis based on the power adjustment data to obtain adjustable load priority data and interrupted load impact priority data;

[0036] A demand smoothing analysis unit, configured to perform demand smoothing analysis based on peak load constraints and valley filling scheduling strategies during valley power periods, using the adjustable load priority data and interrupted load impact priority data to obtain demand smoothing analysis data.

[0037] Optionally, the energy distribution module includes:

[0038] A target function unit, configured to set an energy management target function based on the demand power and the power decision vector;

[0039] A simulation analysis unit, configured to perform a simulation analysis on the charge and discharge of the energy storage system based on the demand smoothing analysis data and the real-time demand in combination with the demand management target function to obtain a simulation analysis result;

[0040] An improvement point unit, configured to extract simulation feedback data based on the simulation analysis result and determine improvement point data based on the simulation feedback data;

[0041] A strategy construction unit, configured to construct an energy management strategy for the energy storage system based on the demand smoothing analysis data and the real-time demand in combination with the energy management target function and the improvement point data.

[0042] In an embodiment of the present invention, the energy storage management system integrates a weather prediction module, an electricity consumption analysis module, a demand analysis module, a demand smoothing module, and an energy distribution module. The weather prediction module is configured to perform weather change prediction based on a weather change prediction model using target historical weather data obtained by encoding preprocessed historical weather data in a latent space, which can improve the accuracy of weather change prediction while eliminating data redundancy. The electricity consumption analysis module is configured to perform an analysis on the changing trend of the electricity consumption load by combining the fuzzy clustering analysis result obtained from the preprocessed historical electricity data with the electricity consumption load prediction curve, which can improve the reliability of the analysis on the changing trend of the electricity consumption load and avoid a large deviation between the obtained electricity consumption load changing trend data and the actual situation. The demand smoothing module is configured to perform demand smoothing analysis based on the real-time demand, the weather change prediction data, and the electricity consumption load changing trend data, which can effectively control the electricity demand, improve the capacity of the energy storage system and the controllable load, and reduce unnecessary electricity consumption. The energy distribution module is configured to determine an energy management strategy for the energy storage system based on the demand smoothing analysis data and the real-time demand in combination with the demand management target function, realize a better energy distribution of the energy storage device, and can avoid excessive charging or discharging of the energy storage device, reduce the peak power, and lower the electricity cost. Description of the Drawings

[0043] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0044] Figure 1 It is a schematic structural composition diagram of an energy storage management system based on cloud deployment control in an embodiment of the present invention;

[0045] Figure 2 It is a schematic flowchart of the energy storage management method based on cloud allocation control in the embodiments of the present invention. Specific Embodiments

[0046] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0047] Embodiment 1

[0048] Please refer to Figure 1 , Figure 1 It is a schematic diagram of the structural composition of the energy storage management system based on cloud allocation control in the embodiments of the present invention. The energy storage management system includes: a pretreatment module 11, a weather prediction module 12, a power consumption analysis module 13, a demand analysis module 14, a demand smoothing module 15, and an energy distribution module 16.

[0049] The pretreatment module 11 is used to preprocess the historical power data and historical weather data of the target power consumption object based on the cloud management system to obtain the preprocessed historical power data and historical weather data;

[0050] In the specific implementation process of the present invention, the pretreatment module 11 is used to perform data cleaning processing and data integration processing on the historical power data and historical weather data to obtain the preprocessed historical power data and historical weather data.

[0051] Specifically, based on the cloud management system, the historical power data of the target power consumption object and the historical weather data of the location where the target power consumption object is located are obtained. The cloud management system is used to obtain data and perform data analysis. The target object can be an industrial and commercial enterprise or a household user, etc. The historical power data includes the electricity price, peak-valley electricity price difference, capacity of the energy storage system, charge and discharge power, efficiency, power consumption load, and load demand in each past time period, etc. The historical weather data includes the weather type data in each past time period. The historical power data and historical weather data are subjected to data cleaning processing, and the abnormal characters and garbled codes in the historical power data and historical weather data are cleaned to obtain the historical power data and historical weather data after data cleaning processing. The historical power data and historical weather data after data cleaning processing are subjected to data integration processing, that is, the data format of the historical power data and historical weather data after data cleaning processing is standardized and the data is classified to obtain the preprocessed historical power data and historical weather data.

[0052] A weather prediction module 12 for predicting weather changes based on preprocessed historical weather data to obtain weather change prediction data;

[0053] In the specific implementation process of the present invention, the weather prediction module 12 includes: a data encoding unit, a feature unit, a model construction unit, and a change prediction unit; the data encoding unit is used to encode the preprocessed historical weather data in the latent space to obtain target historical weather data; the feature unit is used to generate meteorological element features based on historical meteorological data; the model construction unit is used to construct a weather change prediction model based on the meteorological element features and the weather label data of each observation point; the change prediction unit is used to predict weather changes using the target historical weather data based on the weather change prediction model to obtain weather change prediction data.

[0054] Further, the feature unit includes: a grid sub-unit and a feature acquisition sub-unit; the grid sub-unit is used to perform spatial grid processing on the historical meteorological data of each observation point to obtain the spatially grid-processed historical meteorological data; the feature acquisition sub-unit performs feature compression and activation processing on the spatially grid-processed historical meteorological data to obtain corresponding weight coefficients, and generates meteorological element features based on the weight coefficients in combination with the spatially grid-processed historical meteorological data.

[0055] Specifically, since weather changes can affect the electrical energy storage and release of the energy storage system, it is necessary to predict weather changes for subsequent demand analysis. The data encoding unit is used to encode the preprocessed historical weather data in the latent space. The latent space is where the encoder maps the input data to the latent representation space. In this latent representation space, the data is converted into a low-dimensional form, thereby capturing the main features of the input data. Encoding in the latent space means converting the preprocessed historical weather data into low-dimensional data, which can reduce the spatio-temporal dimension of the historical weather data and enhance the feature dimension of the historical weather data. Among them, when encoding in the latent space, the feature dimension of the preprocessed historical weather data is increased through several encoding residual convolutional layers, which can not only reduce data redundancy but also capture important feature data, resulting in the encoded historical weather data, that is, obtaining the target historical weather data. The gridification subunit is used to obtain the historical meteorological data of each observation point. The observation point is the location set by the user for weather phenomenon prediction. The historical meteorological data includes the atmospheric motion change data and atmospheric physical state change data of each observation point at each time period, and performs spatial gridification processing on the historical meteorological data. According to the preset geographical location interval, grids are divided based on the positions of each observation point. Each grid corresponds to the historical meteorological data at the same position in the historical time point, obtaining the historical meteorological data after spatial gridification processing. The feature acquisition subunit performs feature compression and activation processing on the historical meteorological data after spatial gridification processing. It performs convolutional processing on the historical meteorological data after spatial gridification processing, compresses the feature of the historical meteorological data after convolutional processing through a preset spatial dimension, performs an activation operation on the historical meteorological data after feature compression, obtains the corresponding weight coefficient, which is used to characterize the importance of meteorological elements, and generates meteorological element features based on the weight coefficient and the historical meteorological data after spatial gridification processing. The meteorological elements are weighted based on the corresponding weight coefficient in the historical meteorological data, and the meteorological elements are the atmospheric observation values included in the historical meteorological data, obtaining meteorological element features. The model construction unit is used to obtain the weather label data of each observation point. The weather label data is the data configured according to the actual weather phenomena that occur. If the weather phenomenon to be predicted occurs, the weather label data is set to a preset threshold. Based on the weather label data and the meteorological element features, a weather change prediction model is constructed. The initial weather change prediction model is trained through the weather label data and the meteorological element features. The obtained initial weather change prediction model can be a deep convolutional neural network, obtaining the trained weather change prediction model, that is, successfully constructing the weather change prediction model.The change prediction unit is used to perform weather change prediction based on the weather change prediction model using the target historical weather data. The target historical weather data is input into the weather change prediction model to perform weather change prediction for the current time period and the next time period, and preliminary weather change prediction data is obtained. The preliminary weather change prediction data is decoded by a decoder, that is, the feature dimension of the preliminary weather change prediction data is reduced and the spatio-temporal dimension is increased through a number of decoding residual convolutional layers. The number of layers of the decoding residual convolutional layer is the same as that of the encoding residual convolutional layer. Thus, while improving the accuracy of weather change prediction, the prediction speed can be accelerated, and weather change prediction data is obtained.

[0056] The power consumption analysis module 13 is used to analyze the change trend of the power consumption load based on the preprocessed historical power data, and obtain the change trend data of the power consumption load;

[0057] In the specific implementation process of the present invention, the power consumption analysis module 13 includes: a fuzzy clustering unit and a change trend unit; the fuzzy clustering unit is used to perform fuzzy clustering analysis on the preprocessed historical power data to obtain a fuzzy clustering analysis result; the change trend unit is used to obtain a power consumption load prediction curve, and perform power consumption load change trend analysis based on the fuzzy clustering analysis result using the power consumption load prediction curve to obtain the power consumption load change trend data.

[0058] Further, the fuzzy clustering unit includes: a matrix generation sub-unit and a result acquisition sub-unit; the matrix generation sub-unit is used to generate a historical data vector according to the preprocessed historical power data and determine the fuzzy grouping matrix of the historical data vector; the result acquisition sub-unit performs target membership degree analysis and target grouping feature analysis on the historical data vector according to the fuzzy grouping matrix to obtain the target membership degree and the target grouping feature, and performs fuzzy clustering analysis based on the target membership degree and the target grouping feature to obtain the fuzzy clustering analysis result.

[0059] Specifically, the matrix generation subunit is used to generate a historical data vector based on the preprocessed historical power data, determine the number of groups to be divided according to the fuzzy clustering parameters, generate a historical data vector according to the number of groups to be divided, determine the fuzzy grouping matrix of the historical data vector, determine the fuzzy grouping matrix according to the historical data vector and its cost function, obtain a preset grouping, the preset grouping includes a hard grouping and a fuzzy grouping, calculate the variance matrix of the historical data vector, calculate the weighted Euclidean distance between the historical data vector and the features of the preset grouping according to the variance matrix, and construct a cost function according to the weighted Euclidean distance. The result acquisition subunit performs target membership degree analysis and target grouping feature analysis on the historical data vector according to the fuzzy grouping matrix, obtains the target membership degree and the target grouping features, performs fuzzy clustering analysis based on the target membership degree and the target grouping features, obtains the fuzzy clustering analysis result, and correspondingly groups the target membership degree and the target grouping features into target classifications, where the deviation between the historical power data in the target classification and the fuzzy clustering center value is less than a preset threshold, and the number of historical power data in the target classification is greater than the number of historical power data of the power consumption load in other classifications generated by the fuzzy clustering, so as to maximize the mining of power consumption patterns, that is, obtain the fuzzy clustering analysis result. The change trend unit is used to obtain the power consumption load prediction curve, perform curve fitting on the power consumption load data and the power generation power data of each past period to form a power consumption load prediction curve corresponding to each period, perform power consumption load change trend analysis based on the fuzzy clustering analysis result using the power consumption load prediction curve, determine training data according to the fuzzy clustering analysis result and the power consumption load prediction curve in combination with a sliding window, and perform the power consumption load change trend of the current time period and the next time period through the training data and a preset deep learning algorithm to obtain the power consumption load change trend data.

[0060] The demand analysis module 14 is used to determine the real-time demand of the current time period based on the power consumption system topology diagram of the target power consumption object;

[0061] In the specific implementation process of the present invention, the demand analysis module 14 includes: an influence analysis unit, a demarcation point unit, and a real-time demand unit; the influence analysis unit is used to obtain the power change rule based on the preprocessed historical power data, and perform demand prediction influence analysis based on the power change rule to obtain demand prediction influence analysis data; the demarcation point unit is used to determine the power consumption property demarcation point based on the power consumption system topology diagram; the real-time demand unit is used to determine the real-time demand of the current time period based on the power consumption property demarcation point by combining the demand prediction influence analysis data with a linear programming model.

[0062] Further, the real-time demand unit includes: a first real-time demand sub-unit and a demand calculation sub-unit; the first real-time demand sub-unit is used to determine the first real-time demand of the current period according to the electricity property demarcation point in combination with the slip method; the demand calculation sub-unit is used to determine the second real-time demand according to the demand prediction impact analysis data in combination with the linear programming model, and calculate the real-time demand of the current period according to the first real-time demand and the second real-time demand.

[0063] Specifically, the impact analysis unit is used to obtain the power change law based on the preprocessed historical power data, obtain the power change curve of the energy storage system according to the preprocessed historical power data, perform curve change characteristic analysis through the power change curve to obtain the change law of the power change curve, that is, obtain the power change law, and perform demand prediction impact analysis based on the power change law, analyze the factors affecting the demand prediction according to the power change law to obtain the factors affecting the demand prediction, including the magnitude of the power load and the power load fluctuation, etc., and obtain the demand prediction impact analysis data. The demarcation point unit is used to determine the electricity property demarcation point based on the electricity system topology diagram, which has been set according to the structure diagram of the electricity system. The electricity property demarcation point refers to the demarcation point where the maintenance and management scope of the power supply side and the power consumption side in the power system is divided according to the property rights. The first real-time demand sub-unit is used to determine the first real-time demand of the current period according to the electricity property demarcation point in combination with the slip method. The demand calculation sub-unit is used to determine the second real-time demand according to the demand prediction impact analysis data in combination with the linear programming model. The linear programming model is constructed by the maximum charge and discharge power of the energy storage system, the power consumption of the target power consumption object, and the electricity price. Calculate the real-time demand of the current period according to the first real-time demand and the second real-time demand, and calculate the real-time demand of the current period according to the first real-time demand, the second real-time demand and their corresponding weight coefficients to obtain the final real-time demand of the current period. Thus, a more accurate real-time demand can be obtained.

[0064] The demand smoothing module 15 is used to perform demand smoothing analysis based on the real-time demand, weather change prediction data and power consumption load change trend data to obtain demand smoothing analysis data;

[0065] In the specific implementation process of the present invention, the demand flattening module 15 includes: an adjustment data unit, a priority analysis unit, and a demand flattening analysis unit; the adjustment data unit is configured to determine power adjustment data based on the real-time demand, weather change prediction data, and power load change trend data; the priority analysis unit is configured to perform adjustable load priority analysis and interrupted load impact priority analysis based on the power adjustment data to obtain adjustable load priority data and interrupted load impact priority data; the demand flattening analysis unit is configured to perform demand flattening analysis using the adjustable load priority data and interrupted load impact priority data based on peak load constraints and valley filling scheduling strategies during valley power periods to obtain demand flattening analysis data.

[0066] Specifically, the adjustment data unit is used to determine power adjustment data based on the real-time demand, weather change prediction data, and power consumption load change trend data, and determine the installed capacity of the energy storage system, the capacity of adjustable flexible loads, and the capacity of interruptible loads according to the real-time demand, weather change prediction data, and power consumption load change trend data, that is, determine power adjustment data. The adjustable flexible load is a device whose power consumption can be adjusted within a certain range, and the interruptible load is a power load that can be temporarily interrupted under certain circumstances. The priority analysis unit is used to perform priority analysis of adjustable loads and priority analysis of the impact of interruptible loads based on the power adjustment data, that is, determine the factors affecting the priority of adjustable loads through the power adjustment data, including cost-benefit, impact on system stability, and response speed, allocate weights to these factors, determine the scores of adjustable loads on each factor according to the allocated weights, and thus determine the priority of each adjustable load, that is, obtain adjustable load priority data. The priority analysis of the impact of interruptible loads is the same as the priority analysis of adjustable loads. The factors affecting the priority of interruptible loads include the importance of the load, interruption cost, and load type. Obtain adjustable load priority data and interruptible load impact priority data, obtain adjustable load priority data and interruptible load impact priority data. The demand flattening analysis unit is used to perform demand flattening analysis using the adjustable load priority data and interruptible load impact priority data based on the peak load constraint and valley-fill scheduling strategy during valley hours. The peak load constraint is to set the peak load, and the net load after energy storage allocation for the target electricity consumption object cannot exceed the peak load. The valley-fill scheduling strategy during valley hours is used for valley-fill scheduling during valley hours. Determine the energy storage charging power during valley electricity price periods according to the maximum valley-fill power. Determine the discharge power of the energy storage system, the operating power of adjustable flexible loads, the interruption time of interruptible loads, the capacity limit of the system, the response time requirement, and the power demand constraint using the adjustable load priority data and interruptible load impact priority data based on the peak load constraint and valley-fill scheduling strategy during valley hours, that is, obtain demand flattening analysis data. Through demand flattening, the fluctuation and imbalance of power loads can be reduced, the operating pressure of the power grid can be reduced, unnecessary resource waste can be reduced, and at the same time, the electricity cost of the electricity consumption object can be reduced, and the economic benefits of the power system can be improved.

[0067] The energy distribution module 16 is used to determine an energy management strategy based on the demand flattening analysis data and the real-time demand, and control the energy distribution of the energy storage device based on the energy management strategy.

[0068] In the specific implementation process of the present invention, the energy distribution module 16 includes: an objective function unit, a simulation analysis unit, an improvement point unit, and a strategy construction unit; the objective function unit is used to set an energy management objective function based on the demand power and the power decision vector; the simulation analysis unit is used to perform a simulation analysis of the charge and discharge of the energy storage system by combining the demand smoothing analysis data and the real-time demand with the demand management objective function to obtain a simulation analysis result; the improvement point unit is used to extract simulation feedback data based on the simulation analysis result and determine improvement point data based on the simulation feedback data; the strategy construction unit is used to construct an energy management strategy by combining the demand smoothing analysis data and the real-time demand with the energy management objective function and the improvement point data.

[0069] Specifically, the objective function unit is used to set an energy management objective function based on the demand power and the power decision vector, and set the energy management objective function by using the decision vectors of the demand power, the charging power, the discharging power, the price interval time, the demand price, and the optimized demand power with a preset function template. The simulation analysis unit is used to perform a simulation analysis of the charge and discharge of the energy storage system by combining the demand smoothing analysis data and the real-time demand with the demand management objective function, and input the demand smoothing analysis data, the real-time demand, and the demand management objective function into simulation software to perform a simulation analysis of the charge and discharge of the energy storage system to obtain a simulation analysis result. The improvement point unit is used to extract simulation feedback data based on the simulation analysis result, extract the charge and discharge time nodes and the output power of the energy storage system in each time period from the simulation analysis result, and determine improvement point data based on the simulation feedback data, evaluate the simulation feedback data by using a preset expert rule, and determine the improvement point data according to the evaluation result. The strategy construction unit is used to construct an energy management strategy for the energy storage system by combining the demand smoothing analysis data and the real-time demand with the energy management objective function and the improvement point data, adjust the demand smoothing analysis data, the real-time demand, and the energy management objective function by using the improvement point data, determine the output power, the charge and discharge time, and the energy distribution demand of the energy storage device according to the adjusted data, that is, obtain the corresponding energy management strategy, and control the energy distribution of the energy storage device based on the energy management strategy.

[0070] In the embodiment of the present invention, the energy storage management system integrates a weather prediction module, a power consumption analysis module, a demand analysis module, a demand flattening module, and an energy distribution module. The weather prediction module is used to predict weather changes based on a weather change prediction model using target historical weather data obtained by encoding historical weather data after preprocessing in a latent space, which can improve the accuracy of weather change prediction while eliminating data redundancy. The power consumption analysis module is used to analyze the changing trend of the power consumption load by combining the fuzzy clustering analysis result obtained from the historical power data after preprocessing with the power consumption load prediction curve, which can improve the reliability of the analysis of the changing trend of the power consumption load and avoid a large deviation between the obtained power consumption load change trend data and the actual situation. The demand flattening module is used to perform demand flattening analysis based on real-time demand, weather change prediction data, and power consumption load change trend data, which can effectively control power consumption demand, improve the capacity of the energy storage system and controllable loads, and reduce unnecessary power consumption. The energy distribution module is used to determine the energy management strategy of the energy storage system based on the demand flattening analysis data and real-time demand in combination with the demand management objective function, realizing a better energy distribution of the energy storage device, which can avoid excessive charging or discharging of the energy storage device, reduce peak power, and lower the power consumption cost.

[0071] Embodiment 2

[0072] The energy storage management method based on cloud deployment control involved in the second embodiment of the present invention is implemented based on the energy storage management system described in the first embodiment, as Figure 2 shown Figure 2 is a schematic flowchart of the energy storage management method in the embodiment of the present invention, and the method includes:

[0073] S21: Obtain the historical power data of the target power consumption object and the historical weather data of the location where the target power consumption object is located based on the cloud management system, and preprocess the historical power data and historical weather data to obtain the preprocessed historical power data and historical weather data;

[0074] In the specific implementation process of the present invention, historical power consumption data of the target power consumption object and historical weather data of the location where the target power consumption object is located are obtained based on the cloud management system. The cloud management system is used to obtain data and perform data analysis. The target object can be an industrial and commercial enterprise or a household user, etc. The historical power consumption data includes electricity prices, peak-valley electricity price differences, capacities of energy storage systems, charge-discharge powers, efficiencies, power loads, and load demands in each past time period, etc. The historical weather data includes weather type data in each past time period. The historical power consumption data and the historical weather data are subjected to data cleaning processing, that is, abnormal characters and garbled codes in the historical power consumption data and the historical weather data are cleaned to obtain the historical power consumption data and the historical weather data after data cleaning processing. The historical power consumption data and the historical weather data after data cleaning processing are subjected to data integration processing, that is, the data formats of the historical power consumption data and the historical weather data after data cleaning processing are standardized and the data are classified to obtain the historical power consumption data and the historical weather data after preprocessing.

[0075] S22: Encode the preprocessed historical weather data in the latent space to obtain the target historical weather data;

[0076] In the specific implementation process of the present invention, since weather changes will affect the electrical energy storage and release of the energy storage system, it is necessary to predict weather changes for subsequent demand analysis. The preprocessed historical weather data is encoded in the latent space. The latent space is a space where the encoder maps the input data to the latent representation space. In this latent representation space, the data is converted into a low-dimensional form, so that the main features of the input data can be captured. Encoding in the latent space means converting the preprocessed historical weather data into data in a low-dimensional form, which can reduce the spatio-temporal dimension of the historical weather data and enhance the feature dimension of the historical weather data. Among them, when encoding in the latent space, the feature dimension of the preprocessed historical weather data is increased through several encoding residual convolutional layers, which can not only reduce data redundancy but also capture important feature data to obtain the encoded historical weather data, that is, the target historical weather data.

[0077] S23: Generate meteorological element features based on historical meteorological data;

[0078] In the specific implementation process of the present invention, historical meteorological data of each observation point is obtained. The observation point is the location set by the user for weather phenomenon prediction. The historical meteorological data includes the atmospheric motion change data and the atmospheric physical state change data of each time period at each observation point, and the historical meteorological data is subjected to spatial grid processing. According to the preset geographical location interval, grids are divided based on the positions of each observation point. Each grid corresponds to the historical meteorological data at the same position in the historical time point, and the historical meteorological data after spatial grid processing is obtained. Feature compression and activation processing are performed on the historical meteorological data after spatial grid processing. Convolution processing is performed on the historical meteorological data after spatial grid processing. Feature compression is performed on the historical meteorological data after convolution processing through a preset spatial dimension. An activation operation is performed on the historical meteorological data after feature compression to obtain the corresponding weight coefficient. This weight coefficient is used to characterize the importance of meteorological elements, and meteorological element features are generated based on the weight coefficient and the historical meteorological data after spatial grid processing. The meteorological elements in the historical meteorological data after spatial grid processing are weighted according to the corresponding weight coefficient. The meteorological elements are the atmospheric observation values included in the historical meteorological data, and meteorological element features are obtained.

[0079] S24: Construct a weather change prediction model based on the meteorological element features and the weather label data of each observation point;

[0080] In the specific implementation process of the present invention, the weather label data of each observation point is obtained. The weather label data is the data configured according to the actually occurring weather phenomena. If the weather phenomenon to be predicted occurs, the weather label data is set to a preset threshold. A weather change prediction model is constructed based on the weather label data and the meteorological element features. The initial weather change prediction model is trained through the weather label data and the meteorological element features. The initial weather change prediction model can be a deep convolutional neural network, and the trained weather change prediction model is obtained, that is, the weather change prediction model is successfully constructed.

[0081] S25: Use the target historical weather data to perform weather change prediction based on the weather change prediction model, and obtain weather change prediction data;

[0082] In the specific implementation process of the present invention, the target historical weather data is input into the weather change prediction model to perform weather change prediction for the current time period and the next time period, and preliminary weather change prediction data is obtained. The preliminary weather change prediction data is decoded through a decoder, that is, the feature dimension of the preliminary weather change prediction data is reduced and the spatio-temporal dimension is increased through several decoding residual convolutional layers. The number of layers of the decoding residual convolutional layer is the same as that of the encoding residual convolutional layer. Thus, the accuracy of weather change prediction can be improved while the prediction speed is accelerated, and weather change prediction data is obtained.

[0083] S26: Perform fuzzy clustering analysis on the preprocessed historical power data to obtain the fuzzy clustering analysis result;

[0084] In the specific implementation process of the present invention, generate a historical data vector according to the preprocessed historical power data, determine the number of groups to be divided according to the fuzzy clustering parameters, generate a historical data vector according to the number of groups to be divided, determine the fuzzy grouping matrix of the historical data vector, determine the fuzzy grouping matrix according to the historical data vector and its cost function, obtain the preset grouping, the preset grouping includes hard grouping and fuzzy grouping, calculate the variance matrix of the historical data vector, calculate the weighted Euclidean distance between the historical data vector and the features of the preset grouping according to the variance matrix, and construct a cost function according to the weighted Euclidean distance. Perform target membership degree analysis and target grouping feature analysis on the historical data vector according to the fuzzy grouping matrix to obtain the target membership degree and target grouping features, perform fuzzy clustering analysis based on the target membership degree and target grouping features to obtain the fuzzy clustering analysis result, and correspondingly group the target membership degree and target grouping features into target classifications, where the deviation between the historical power data in the target classification and the fuzzy clustering center value is less than the preset threshold, and the number of historical power data in the target classification is greater than the number of historical power data of the electricity consumption load in other classifications generated by the fuzzy clustering, and the electricity consumption pattern can be mined to the greatest extent, that is, the fuzzy clustering analysis result is obtained.

[0085] S27: Based on the fuzzy clustering analysis result, use the electricity consumption load prediction curve to perform electricity consumption load change trend analysis to obtain the electricity consumption load change trend data;

[0086] In the specific implementation process of the present invention, obtain the electricity consumption load prediction curve, perform curve fitting on the electricity consumption load data and power generation power data of each past period to form an electricity consumption load prediction curve corresponding to each period, and use the electricity consumption load prediction curve based on the fuzzy clustering analysis result to perform electricity consumption load change trend analysis. Determine the training data according to the fuzzy clustering analysis result and the electricity consumption load prediction curve in combination with a sliding window, and perform the electricity consumption load change trend of the current time period and the next time period through the training data and the deep learning algorithm of the preset value to obtain the electricity consumption load change trend data.

[0087] S28: Determine the real-time demand of the current time period based on the power system topology diagram of the target electricity consumption object;

[0088] In the specific implementation process of the present invention, the power change law is obtained based on the preprocessed historical power data, the power change curve of the energy storage system is obtained according to the preprocessed historical power data, the curve change characteristics analysis is carried out through the power change curve to obtain the change law of the power change curve, that is, the power change law is obtained, and the impact analysis of demand prediction is carried out based on the power change law, the factor analysis affecting the demand prediction is carried out according to the power change law to obtain the factors affecting the demand prediction, which includes the size of the power load and the power load fluctuation, etc., and the impact analysis data of demand prediction is obtained. The power consumption property demarcation point is determined based on the power consumption system topology diagram, and the power consumption system topology diagram has been set according to the structure diagram of the power consumption system. The power consumption property demarcation point refers to the demarcation point where the maintenance and management scopes of the power supply side and the power consumption side in the power system are divided according to the property rights. The first real-time demand of the current period is determined according to the power consumption property demarcation point in combination with the slip method. The second real-time demand is determined according to the impact analysis data of demand prediction in combination with the linear programming model. The linear programming model is constructed by the maximum charge and discharge power of the energy storage system, the power consumption of the target power consumption object and the electricity price. The real-time demand of the current period is calculated according to the first real-time demand and the second real-time demand. The real-time demand of the current period can be calculated according to the first real-time demand, the second real-time demand and their corresponding weight coefficients to obtain the final real-time demand of the current period. Thus, a more accurate real-time demand can be obtained.

[0089] S29: Carry out demand smoothing analysis based on the real-time demand, weather change prediction data and power consumption load change trend data to obtain demand smoothing analysis data;

[0090] In the specific implementation process of the present invention, power adjustment data is determined based on the real-time demand, weather change prediction data, and power load change trend data. The installed capacity of the energy storage system, the capacity of adjustable flexible loads, and the capacity of interruptible loads are determined according to the real-time demand, weather change prediction data, and power load change trend data, that is, power adjustment data is determined. An adjustable flexible load is a device whose power consumption can be adjusted within a certain range, and an interruptible load is a power load that can be temporarily interrupted under certain circumstances. Based on the power adjustment data, an analysis of the priority of adjustable loads and an analysis of the priority of the impact of interruptible loads are carried out. That is, the factors affecting the priority of adjustable loads are determined through the power adjustment data, including cost-benefit, impact on system stability, and response speed. Weights are assigned to these factors, and the scores of each adjustable load on each factor are determined according to the assigned weights to determine the priority of each adjustable load, that is, the priority data of adjustable loads is obtained. The steps of the analysis of the priority of the impact of interruptible loads are the same as those of the analysis of the priority of adjustable loads. The factors affecting the priority of interruptible loads include the importance of the load, interruption cost, and load type. The priority data of adjustable loads and the priority data of the impact of interruptible loads are obtained. Based on the peak load constraint and the valley filling scheduling strategy during off-peak hours, the demand flattening analysis is carried out using the priority data of adjustable loads and the priority data of the impact of interruptible loads. The peak load constraint means setting a peak load, and the net load after energy storage allocation for the target power consumption object cannot exceed the peak load. The valley filling scheduling strategy during off-peak hours is used for valley filling scheduling during off-peak hours. The energy storage charging power during the valley electricity price period is determined according to the highest valley filling power. According to the peak load constraint and the valley filling scheduling strategy during off-peak hours, the discharge power of the energy storage system, the operating power of adjustable flexible loads, the interruption time of interruptible loads, the capacity limit of the system, the response time requirement, and the power demand constraint are determined using the priority data of adjustable loads and the priority data of the impact of interruptible loads, that is, the demand flattening analysis data is obtained. Through demand flattening, the fluctuations and imbalances of power loads can be reduced, the operating pressure of the power grid can be lowered, unnecessary resource waste can be reduced, and at the same time, the power cost of the power consumption object can be reduced, and the economic benefits of the power system can be improved.

[0091] S210: Determine an energy management strategy based on the demand flattening analysis data and the real-time demand in combination with the demand management objective function, and control the energy distribution of the energy storage device based on the energy management strategy.

[0092] In the specific implementation process of the present invention, an energy management objective function is set based on the demand power and the power decision vector. The energy management objective function is set by using a preset function template with the decision vectors of the demand power, the charging power, the discharging power, the price interval time, the demand price, and the optimized demand power. Based on the demand flattening analysis data and the real-time demand, combined with the demand management objective function, a simulation analysis of the charge and discharge of the energy storage system is carried out. The demand flattening analysis data, the real-time demand, and the demand management objective function are input into the simulation software for the simulation analysis of the charge and discharge of the energy storage system, and the simulation analysis results are obtained. Based on the simulation analysis results, simulation feedback data is extracted. The charge and discharge time nodes and the output power of the energy storage system in each time period are extracted from the simulation analysis results, and improvement point data is determined based on the simulation feedback data. The simulation feedback data is evaluated by preset expert rules, and the improvement point data is determined according to the evaluation results. Based on the demand flattening analysis data and the real-time demand, combined with the energy management objective function and the improvement point data, an energy management strategy for the energy storage system is constructed. The demand flattening analysis data, the real-time demand, and the energy management objective function are adjusted by the improvement point data, and the output power, the charge and discharge time, and the energy distribution demand of the energy storage device are determined according to the adjusted data, that is, the corresponding energy management strategy is obtained. Based on the energy management strategy, the energy distribution of the energy storage device is controlled.

[0093] In the embodiment of the present invention, based on the weather change prediction model, the target historical weather data obtained by encoding the preprocessed historical weather data in the hidden space is used for weather change prediction, which can improve the accuracy of weather change prediction while eliminating data redundancy. By combining the fuzzy clustering analysis results obtained from the preprocessed historical electrical energy data with the electricity load prediction curve, the analysis of the change trend of the electricity load is carried out, which can improve the reliability of the analysis of the change trend of the electricity load and avoid the deviation of the obtained electricity load change trend data from the actual situation being too large. Based on the real-time demand, the weather change prediction data, and the electricity load change trend data, the demand flattening analysis is carried out, which can effectively control the electricity demand, improve the capacity of the energy storage system and the controllable load, and reduce unnecessary electricity consumption. Based on the demand flattening analysis data and the real-time demand, combined with the demand management objective function, the energy management strategy of the energy storage system is determined, and the energy distribution of the energy storage device with better effect is realized, which can avoid excessive charging or discharging of the energy storage device, reduce the peak power, and reduce the electricity cost.

[0094] Those of ordinary skill in the art can understand that all or part of the steps in the above methods of the embodiments can be completed by instructing relevant hardware through a program, and the program can be stored in a computer-readable storage medium. The storage medium can include: read-only memory (ROM, Read Only Memory), random access memory (RAM, Random Access Memory), magnetic disk or optical disk, etc.

[0095] In addition, the above has introduced in detail an energy storage management system based on cloud deployment control provided by the embodiments of the present invention. Specific examples should have been used in this article to elaborate on the principle and implementation manner of the present invention. The description of the above embodiments is only used to help understand the method and its core idea of the present invention; at the same time, for those of ordinary skill in the art, according to the idea of the present invention, there will be changes in the specific implementation manner and application scope. In summary, the content of this specification should not be construed as a limitation to the present invention.

Claims

1. An energy storage management system based on cloud deployment control, characterized in that, The energy storage management system includes: A preprocessing module for preprocessing historical power data and historical weather data of a target power consumption object based on a cloud management system to obtain preprocessed historical power data and historical weather data; A weather prediction module for predicting weather changes based on the preprocessed historical weather data to obtain weather change prediction data; A power consumption analysis module for analyzing the changing trend of power consumption load based on the preprocessed historical power data to obtain power consumption load changing trend data; A demand analysis module for determining the real-time demand in the current period based on the power consumption system topology diagram of the target power consumption object; A demand flattening module for performing demand flattening analysis based on the real-time demand, weather change prediction data, and power consumption load changing trend data to obtain demand flattening analysis data; An energy distribution module for determining an energy management strategy based on the demand flattening analysis data and the real-time demand to control the energy distribution of energy storage devices.

2. The energy storage management system according to claim 1, wherein The preprocessing module is used to perform data cleaning processing and data integration processing on the historical power data and historical weather data to obtain preprocessed historical power data and historical weather data.

3. The energy storage management system according to claim 1, wherein The weather prediction module includes: A data encoding unit for encoding the preprocessed historical weather data in a latent space to obtain target historical weather data; A feature element unit for generating meteorological element features based on historical meteorological data; A model construction unit for constructing a weather change prediction model based on the meteorological element features and weather label data of each observation point; A change prediction unit for predicting weather changes using the target historical weather data based on the weather change prediction model to obtain weather change prediction data.

4. The energy storage management system according to claim 3, wherein The feature element unit includes: A grid division subunit for performing spatial grid division processing on the historical meteorological data of each observation point to obtain spatially grid-divided historical meteorological data; A feature acquisition subunit for performing feature compression and activation processing on the spatially grid-divided historical meteorological data to obtain corresponding weight coefficients, and generating meteorological element features based on the weight coefficients and the spatially grid-divided historical meteorological data.

5. The energy storage management system according to claim 1, wherein The power consumption analysis module includes: A fuzzy clustering unit for performing fuzzy clustering analysis on the preprocessed historical power data to obtain a fuzzy clustering analysis result; A changing trend unit for analyzing the changing trend of power consumption load using a power consumption load prediction curve based on the fuzzy clustering analysis result to obtain power consumption load changing trend data.

6. The energy storage management system according to claim 5, wherein, The fuzzy clustering unit includes: A matrix generation subunit for generating a historical data vector according to the preprocessed historical power data and determining a fuzzy grouping matrix of the historical data vector; A result acquisition subunit for performing target membership degree analysis and target grouping feature analysis on the historical data vector according to the fuzzy grouping matrix to obtain a target membership degree and target grouping features, and performing fuzzy clustering analysis based on the target membership degree and target grouping features to obtain a fuzzy clustering analysis result.

7. The energy storage management system according to claim 1, wherein The demand analysis module includes: An impact analysis unit, configured to obtain the power change law based on the preprocessed historical power data, and perform a demand prediction impact analysis based on the power change law to obtain demand prediction impact analysis data; A demarcation point unit, configured to determine the electricity property demarcation point based on the electricity consumption system topology diagram; A real-time demand unit, configured to determine the real-time demand of the current period by using the demand prediction impact analysis data in combination with a linear programming model based on the electricity property demarcation point.

8. The energy storage management system according to claim 7, characterized in that The real-time demand unit includes: A first real-time demand sub-unit, configured to determine the first real-time demand of the current period by combining the slip method with the electricity property demarcation point; A demand calculation sub-unit, configured to determine the second real-time demand by combining the demand prediction impact analysis data with a linear programming model, and calculate the real-time demand of the current period according to the first real-time demand and the second real-time demand.

9. The energy storage management system according to claim 1, characterized in that The demand smoothing module includes: An adjustment data unit, configured to determine power adjustment data based on the real-time demand, weather change prediction data, and electricity load change trend data; A priority analysis unit, configured to perform an adjustable load priority analysis and an interrupted load impact priority analysis based on the power adjustment data to obtain adjustable load priority data and interrupted load impact priority data; A demand smoothing analysis unit, configured to perform a demand smoothing analysis by using the adjustable load priority data and the interrupted load impact priority data based on the peak load constraint and the valley filling scheduling strategy during the valley electricity period to obtain demand smoothing analysis data.

10. The energy storage management system according to claim 1, characterized in that, The energy distribution module includes: An objective function unit, configured to set an energy management objective function based on the demand power and the power decision vector; A simulation analysis unit, configured to perform a simulation analysis of the charge and discharge of the energy storage system by combining the demand smoothing analysis data and the real-time demand with the demand management objective function to obtain a simulation analysis result; An improvement point unit, configured to extract simulation feedback data based on the simulation analysis result, and determine improvement point data based on the simulation feedback data; A strategy construction unit, configured to construct an energy management strategy by combining the demand smoothing analysis data and the real-time demand with the energy management objective function and the improvement point data.

Images